Passkeys still provide valuable protection against credential phishing, but deploying them does not finish an account-security project. Anvil Secure's October 1, 2026 research release is a timely reason to examine the systems around sign-in: application checks, credential enrollment, client software, and account access after authentication. Its announcement describes a broad assessment framework, rather than a single newly affected product with a universal patch. Anvil Secure announcement

This analysis was prepared on October 2, 2026. The research publication and announcement are dated October 1. Readers should distinguish that publication date from the dates of individual vulnerabilities discussed in the underlying research.

What Anvil Secure published

Matteo Giordano's 60-page whitepaper, Demystifying Passkeys: Under Attack, examines authenticators and protocols, cross-device transport, clients, relying parties, synchronization, and user recovery. It includes a 43-row testing reference. The authors distinguish research into an attack class from an assessment of a particular deployment. Those are different kinds of evidence: a research possibility alone does not establish that a particular company's accounts are exploitable. Whitepaper

The scope is passkey deployments across multiple components. There is no single affected-version range or overall remediation status for this publication. Any specific vulnerability requires its own advisory, prerequisites, affected versions, and fix assessment. This article does not identify a newly disclosed vulnerability in Thetis products or recommend an emergency hardware replacement.

Ask what the application actually verifies

WebAuthn is the web standard behind passkey authentication. Its server-side verification procedure checks more than a cryptographic signature: the expected challenge, origin, relying-party identifier, and relevant authenticator flags also matter. The implementation must connect the resulting credential to the correct account. W3C WebAuthn verification requirements

Our editorial recommendation is to make an application owner explain that connection in plain language. Which account receives a newly registered credential? Which environment issued the challenge? Which component decides whether a response is accepted? A clear answer identifies where an independent review belongs. “We installed a passkey library” is useful implementation context, but it is not a completed review record.

For an assessment brief, request expected outcomes rather than an indiscriminate collection of attack demonstrations. The reviewer should document the authentication boundary, the controls examined, and any exceptions requiring a business decision. Keep testing within explicitly authorized accounts and systems.

Include what happens after sign-in

A successful authentication typically establishes an application session. OWASP explains that the session token then represents the authenticated access; its theft or misuse can enable impersonation. OWASP recommends controls including session expiration, appropriate cookie protections, and reauthentication after risk events. A passkey should therefore not be described as automatically preventing session hijacking or endpoint malware. OWASP Session Management Cheat Sheet

Consider a hypothetical finance application. Its sign-in page may use passkeys, while a later workflow changes payment details. Our recommendation is to review those as separate decisions: gaining access to the application and authorizing a sensitive change. The organization should decide when a fresh authentication is needed, what event is recorded, and how an operator investigates an unexpected change. This is an illustrative review scenario, not a report of a tested vulnerability.

That approach also gives security teams a more useful success measure. Count reviewed access paths and resolved findings alongside passkey enrollment. Enrollment alone cannot describe the entire protection of a business workflow.

Put hardware keys in the right part of the review

A FIDO2 security key can hold a device-bound passkey. That makes the authenticator choice relevant, but it does not replace application review. The whitepaper explicitly separates different credential arrangements because their trust assumptions differ. Its discussion should not be flattened into a claim that every passkey shares every reported risk. Anvil Secure whitepaper

For a hardware-key evaluation, our recommendation is to ask the service owner for explicit external-key support and test the intended browser, operating system, and connection. A general announcement of passkey availability is insufficient evidence of that compatibility. Include the exact sign-in path in the review record so that a successful browser test is not silently treated as proof about every native application.

Teams considering this option can explore Thetis FIDO2 security keys and compare the relevant model documentation with their service requirements. Product selection should follow the use case and verified compatibility.

Turn the research into one concrete next step

Our suggested next step is a short scoping meeting with the identity owner, application owner, and security reviewer. Select one important application and produce a one-page review brief covering entry points, credential enrollment, sensitive actions, and sessions. Assign an owner and expected evidence to each area.

Keep three statuses separate: reviewed and acceptable, a confirmed finding awaiting remediation, and not yet assessed. That distinction prevents an untested assumption from becoming a reassuring green checkmark. The useful outcome of this week's research is a better review scope and an accountable follow-through process—not a blanket verdict on every passkey implementation.